UniPath

The software tool 'UniPath' addresses challenges associated with analyzing single-cell open-chromatin and transcriptome data that exhibit high variability in noise and drop-out rates among cells. 'UniPath' introduces a transformative method for representing single cells by generating pathway and gene-set enrichment scores based on their open-chromatin or gene-expression profiles. This approach offers robust statistical accuracy, consistency, and scalability in estimating enrichment scores for each cell, while also accommodating variability in drop-out rates that can lead to artifacts. The tool also presents an innovative dimension reduction technique for single-cell open-chromatin profiles. Moreover, 'UniPath' predicts the temporal order of single cells using their pathway enrichment scores, enabling the accurate suppression of covariates and correct cell ordering. The tool's application to a mouse cell atlas yielded biologically meaningful co-clustering of cell types from different organs. Additionally, the tool's utilization of pathway co-occurrence facilitates the comparison of cell groups, making it useful for inferring context-specific regulations in cancer cells.